by Dr. Kendra Zamzow
On January 26th, the Feijao tailings dam in Brazil released 85 percent of the 13 million m3 of iron-ore tailings behind it. At least 84 people died, and the final death toll is likely to be in the hundreds, most company employees (https://www.theglobeandmail.com/world/article-brazil-learned-nothing-another-deadly-dam-collapse-raises-questions/).
German inspectors certified the dam as safe less than a year before, and the dam had monitoring equipment and biweekly inspections (http://www.vale.com/EN/aboutvale/news/Pages/Clarifications-regarding-Dam-I-of-the-Corrego-do-Feijao-Mine.aspx; https://www.cnbc.com/2019/01/29/reuters-america-vale-eyed-dam-design-changes-in-2009-that-may-have-prevented-disaster.html).
In 2015, a similar failure in Brazil released 82% of the 55 million m3 of tailings held, tailings which flowed 250 miles to the sea. In 2014, the Mt. Polley mine in British Columbia released 32% of 74 million m3 of tailings, which ran 4 miles to a salmon lake. In 1998 and 2000, failures released 45% and 12% of the tailings at dams in Sweden and Spain at smaller mines. I do not know what the Feijao dam was built of, but of the 104 tailings dam failures that occurred globally in the last 20 years (https://worldminetailingsfailures.org) the other dams I list above were all made of mine waste or rock, generally strong material. They were also traditional “slurry” tailings impoundments.
Donlin proposes a rock dam holding back tailings slurry. They promised to put the foundation on bedrock, a good start. Donlin agreed to independent inspections. Feijao had independent inspectors – two employees of that firm are now under arrest. Donlin will put monitoring equipment in the dam. Feijao had that, too. Donlin promises to use “downstream” construction. Many dams start as downstream. The “downstream method” can be changed over time, as can the steepness of the downstream face. Mt. Polley was to be a 2 (horizontal) to 1 (vertical) slope, but was much steeper when it failed.
Dams are built in stages as the mine expands. When the dam needs to be raised, dump trucks containing construction rock will roll along the wide crest of the dam and decide – do I dump the dam construction rock a few feet down onto the tailings side, or hundreds of feet down on the face of the dam (“downstream”)? If a mine operator is in a hurry, dumping on the tailings side (“upstream”) increases dam height with less rock, less time, less truck fuel. Feijao used the upstream method. Mt. Polley went “centerline”, dumping some on the tailings side and some on the face, and went upstream over time. Decisions are made at points in the future that coincide with other forces – swings in metal prices, the availability of construction rock, mine sites changing hands.
I participated in cooperating agency meetings for the Donlin mine from 2013-2017. I requested modeling of spills of 5 to 20 percent. Not unreasonable given the failures documented. The model was done for a spill of 0.5 percent of tailings. Donlin Gold noted that they did not have to model “catastrophic” failures. Would a release of more than 0.5 percent of the tailings be catastrophic?
Alaskans are moving away from dangerous slurry impoundments. Pogo and Greens Creek mine use the “dry stack” method. “Paste tailings” suggested in scoping for Donlin were rejected. The dry stack was never a serious alternative in the Environmental Impact Statement (EIS). The Donlin mine will hold 412 million m3 of tailings when fully built – much more than the mines mentioned above. The probability of dam failure at Donlin may be “very low” as stated in the EIS– but impacts would be generational.
Paste and dry stack may pose technical challenges. But where risk is high, if slurry alternatives are not feasible perhaps it is time for a process where “No mine” is a real option. Society has moved from putting tailings in streams to putting them behind dams; it’s time to take the next step in containment.
At the end of mining, Donlin tailings water will be drained into the pit. But water will remain entrained within tailings and the dam itself. Danger is abated, but not eliminated. The degree to which it is abated isn’t known; that wasn’t modeled either.
DNR is reviewing Donlin’s application to construct a dam. DNR could request a full comparison of dry stack, paste, and slurry tailings in terms of economics, technical challenges, and environmental risks. DNR could require modeling of various spill sizes – for emergency preparedness if for no other reason.
The question is – will they?
Dr. Zamzow received her PhD in Environmental Chemistry at the University of Nevada, Reno in 2007. Since 2008 she has worked for the Center for Science in Public Participation (www.csp2.org) assisting communities with evaluating the risks of mining projects. She has lived in Alaska for over 30 years, and currently lives in Chickaloon.